Outline

• Package (15 min)
  • How to find a package on internet
  • Various way to install a package(Cran, Github and more)
  • How to read package manual
• Data I/O(25 min)
  • Tabular input
  • JSON
  • Databases(SQLite)
• magrittr (10 min)
  • Play with magrittr
• data.table (80 min)
  • Concept of data.table
  • How to read data with data.table
  • Slicing, filtering and aggregation
  • Deep copy and shallow copy
  • Sort and merge data
• Schedualing(20 min)
  • Task Scheduler(Windows)
  • Crontab(Linux/Mac)
• R Learning Resource
  • datacamp, tryR
  • Coursera JohnsHopkins, Udacity Data Analysis with R, Nanodegree…
  • Adv.R

libPaths

Libraries in R are loaded in order:

.libPaths()

## [1] "/usr/local/lib/R/site-library" "/usr/lib/R/site-library"      
## [3] "/usr/lib/R/library"

You can also add an additional library path:

.libPaths(c("your/library/path", .libPaths()))

Change Privilege(Windows)

Task Schedualing in R

Sometime you may need to execute your R script periodically and automatically:

• Regular Report: Generate monthly/daily report automatically
• Crawler: Get latest data every hour
• Forecasting System: Rebuild your preditive model with latest data

Steps to Schedualing

1. Invoking a batch job of R
   • RScript
   • R CMD BATCH
2. Passing arguments to R
3. A schedualer:
   • Linux/Mac: Crontab
   • Windows: Task Schedualer

Open shell with RStudio

Environment variable of R and RScript will be exported in this way.

Invoke a batch job

• Rscript
  • A binary front-end to execute a R script
  • Recommended
  • Execute a R script:

Rscript <PATH OF YOUR R SCRIPT> <ARG1> <ARG2> ...

• R CMD BATCH
  • Run R in batch mode
  • I never use this one since I use Rscript

Shebang

If you use Linux or Mac, you can execute a R script directly with shebang and permission of execution.

#!/usr/bin/env Rscript

Add permission of execution:

chmod u+x schedualing/now.R

Check permission:

ls -l schedualing/now.R

## -rwxrw-rw- 1 mansun mansun 1164 May  5 01:24 schedualing/now.R

Passing arguments

• Function:
  • commandArgs: a function to access command line arguments
• Package:
  • argparser: Command-line argument parser written purely in R
  • ArgParser: a command line argument parser for R from Kyle
  • argparse: Analog to Python’s argparse
  • optparse: Analog to Python’s optparse

Exercise

Let’s try to pass some arguments to schedualing/now.R:

Get help:

schedualing/now.R -h

Pass a positional argument:

schedualing/now.R Mansun

Add a flag -c:

schedualing/now.R Mansun -c

Add a optional argument:

schedualing/now.R Mansun -c -g "How are you?"

Crontab(Linux/Mac)

Maintain crontab files to execute scheduled commands in Unix-like OS for individual users.

crontab [-u user] file  
crontab [ -u user ] [ -i ] { -e | -l | -r }  
    -e      (edit user's crontab)  
    -l      (list user's crontab)  
    -r      (delete user's crontab)  
    -i      (prompt before deleting user's crontab)  

you may need to use some command line editor like vim when using crontab -e to edit your crontab.

Your first crontab

• Open your terminal
• Download our github repo

cd; mkdir github; cd github;

• Type select-editor to choose your favorite editor
• Type crontab -e to edit
• Enter following text in your terminal and save:

* * * * * cd /home/mansun/github/BeyondBasicR/schedualing; ./now.R Mansun

• Type crontab -l to check your crontab:

crontab -l

## * * * * * cd /home/mansun/github/BeyondBasicR/schedualing; ./now.R Mansun

┌───────────── min (0 - 59) 
│ ┌────────────── hour (0 - 23)
│ │ ┌─────────────── day of month (1 - 31)
│ │ │ ┌──────────────── month (1 - 12)
│ │ │ │ ┌───────────────── day of week (0 - 6) (0 to 6 are Sunday to Saturday)
│ │ │ │ │
│ │ │ │ │
* * * * *  command to execute

• _*_:

Tables

Task Schedualer(Windows)

Launch task schedualer within RStudio:

# Execute a system conmmand to launch task schedualer
system("control schedtasks")

Samples

Set a task

References

• cron
• Scheduling R Tasks via Windows Task Scheduler
• Managing libraries